Manually created links bug? Or what?

The rendering engine on the site is doing something very odd... at least I think it is the engine that is acting up.

I am currently experimenting with parsing HTML and such, so I needed some examples. I did a "view source" on my home node, and used that to play with. After toying around, I put it through a HTML validator, and it was very upset that I had nested link (<a>) tags. And it turned out that it wasn't my parsing either, the original source from the page has nested, double links.

This is when it happens: When someone creates a normal link, using <a href=.... instead of the internal linking mechanism, which is necessary to create links with query strings. So far I've seen this behaviour on my own, and crazyinsomniac's home nodes (haven't looked further when I found another place), and at least on my node, the HTML is "sane" when I edit it. I have no doubt it is on his too.

So far, I've gotten the same behaviour from IE6 on windows 2000, and I've had opera on linux identify itself as IE5, mozilla 5.0 and Opera 6 so far - completely purging all cache and such between tries. Exact same error. Do a "view source" and search for '</a></a>' and you should see it.

Update: It seems that it matters if I am logged in or not. When logging out, the behaviour disappeared, which explains the lynx/perl issue below. I'm not sure if that doesn't make it *more* strange though...

But wait - it gets better! When I do a fetch of the page with lynx (linux) or LWP::Simple on the command line (both platforms), the HTML is sane!

Since our very openminded browsers seem to render this just fine anyways, it is no urgent issue, but: a) it is definetely some kind of bug, and b) I don't know what other things it might affect, so I thought I'd better bring it up right away.

I really doubt it is possible for both browsers to have this bug in it's rendering, especially since there really isn't anything odd about the original HTML. :)

You have moved into a dark place.
It is pitch black. You are likely to be eaten by a grue.

Comment on Manually created links bug? Or what? Select or Download Code

Replies are listed 'Best First'.
(tye)Re: Manually created links bug? Or what? by tye (Sage) on May 13, 2002 at 14:18 UTC
It sounds like you failed to notice the key fact that these links are in the Personal Nodelet (and so which page you are looking at doesn't make any difference). The Personal Nodelet expects the links you give it to either be all digits (in which case it links by node_id) or to be a node title. However, if you instead give it `<a ...>...</a>`, it produces something that works in most browsers so many people have done that. It sounds like you have done this. At some point, we'll make the personal nodelet smarter and it will no longer try to construct a link to a node with a title of "`<a ...>...</a>`" and then fail to escape the "title" of the link. Instead, it will notice these cases and simple include your link (or whatever HTML you entered). - tye (but my friends call me "Tye")	[reply] [d/l] [select]
Re: (tye)Re: Manually created links bug? Or what? by Dog and Pony (Priest) on May 13, 2002 at 14:29 UTC
D'oh. I'm so stupid. And really sorry to have bothered you with this. Of course that is the answer. I'd try to excuse myself with the facts that it was really late, that I had the exact same links in the page itself, and that the html isn't all that easy to read, since it is a lot and cluttered. But that won't cut it... :) Thank you for the polite and clarifying reply! :) You have moved into a dark place. It is pitch black. You are likely to be eaten by a grue.	[reply]